Research on infrared target recognition algorithm based on SSD-MobileNetv1 network
Siyuan Cheng, Dezhen Yang, Jiangyong Li, Peng Jia · Seventh Asia Pacific Conference on Optics Manufacture and 2021 International Forum of Young Scientists on Advanced Optical Manufacturing (APCOM and YSAOM 2021) · 2022
Infrared target recognition is a hot direction in computer vision and digital image processing, which has high theoretical research value and market application prospects, and develops gradually with the development of deep learning, image processing, pattern recognition and other fields. In order to solve the problem of insufficient accuracy and recognition speed of infrared target recognition in practical application. An infrared target recognition method based on SSD-MobileNetv1 target detection model is constructed. In this paper, the self-made infrared data set is used, and the data set is enhanced and annotated, and the noise is suppressed. In the Windows system, the object detection API of TensorFlow is used to train the data set, and the trained SSD-MobileNetv1 model is used to classify the test images and obtain the recognition results. The experimental results show that: The SSD-MobileNetv1 model has good detection and recognition effects on infrared images; The model has the advantages of accurate target location, high recognition accuracy, and good stability for the recognition of images with background interference.